How will the COVID pandemic affect productivity growth? My long-read Q&A with Chad Syverson

Once we get the COVID–19 crisis under control, we’ll need to support a quick economic recovery. That requires us to know what the economy might look like in the post-pandemic future. So what will the economy look like? In particular, how will the pandemic affect the stagnant productivity growth of the past 15 years? Chad Syverson joined Political Economy this week to discuss.

Chad is the George C. Tiao Distinguished Service Professor of Economics at the University of Chicago’s Booth School of Business. Along with Filippo di Mauro, he recently wrote an article for VoxEU, titled “The COVID crisis and productivity growth”.

What follows is a lightly edited transcript of our conversation, including brief portions that were cut from the original podcast. You can download the episode here, and don’t forget to subscribe to my podcast on iTunes or Stitcher. Tell your friends, leave a review.

Pethokoukis: The analysis that you co-authored looks at how the coronavirus pandemic will affect the medium- and long-term productive capacity of the US economy going forward. But before I get into that, I wanted to just focus a bit on something else you wrote about where we were heading before the crisis. You and your authors co-wrote: “The world entered into the COVID crisis in the midst of a 15-year-long productivity growth slowdown. While much debated, there is not yet a consensus on its causes.”

So there’s no consensus on this productivity slowdown — which, in a way, has actually been going on since maybe the early 1970s. But what’s your best guess?

Syverson: I think in the last 15 years, the US ran out of the low-hanging fruit that the first IT wave brought. Now, there was a clear productivity growth acceleration from the mid-1990s to about 2004 or so, and you could trace that through the economy. First, it was in the production of IT goods, and then it moved into industries that used IT goods intensively. So you can sort of trace out how new ideas were working their way through the economy and yielding productivity growth.

And then as those first, more obvious uses of those technologies were harnessed, you start running into more and more difficulty finding new ways to get productivity gains. I think that is a good candidate for the slowdown of the last, say, 15 years. You’re right that if you take a broader view, with the exception of that acceleration I just mentioned, we are in about a 50-year long period of pretty modest productivity growth.

So 1974 seems to be special. Since that point, we haven’t been able to really sustain — with the exception of the 10-year period I mentioned in the US — average labor productivity growth much above two percent per year. And I don’t have a really sharp sense of if that’s one explanation for that, or if there are multiple things. There are so many candidates.

My guess is anything that big and that long probably is multi-causal. Which things explain it, and how important is each one? That’s a hard question.

Is that a question people are still currently looking at? Is there any consensus on things which are maybe more likely versus somewhat less likely?

Syverson: So, I think Bob Gordon is pretty confident of his answer, which is that the nature of technological change and our ability to interact with it and create more of it has just changed since 1970. We shouldn’t expect to get back to those pre-1974 productivity growth levels again in the long run. I understand the case, but that doesn’t strike me as the right explanation.

I’ve written about this with Erik Brynjolfsson and Daniel Rock: When new technologies come along, there is a long period — we’re talking decades — where the technology’s potential is recognized in some general sense, and people can imagine, “Oh wow, wouldn’t this industry be different if they could figure out how to harness this technology and do this, that, and the other thing?”

But it takes a long time for “this, that, and the other things” to actually happen. A good example that’s really salient right now is eCommerce retail. I’m old enough to remember the 1990s: People saw Amazon and other eCommerce companies and at the time they thought, “Wow, this really might change the way retail is done.”

But even now, eCommerce is probably — depending on how you measure it — 15 percent of total retail sales. It took 25 years to get to that point. And arguably, maybe the crisis will accelerate it a little bit, but it just took a long time for the sector itself and consumers who interact with the sector, to change behavior and processes enough for that eCommerce potential. Which, again, was recognizable 25 years ago to actually start reshaping how things were done.

So I think the most optimistic view of the current slowdown is this: It’s prep work for the next wave.

I’m hoping for an AI productivity boom going forward, but there is just not a lot of evidence of that so far. So the optimistic case is: Don’t rule it out, we could just be still on the “figuring out” part of the process.

Yeah, exactly. That’s the idea.

I love economic growth. I would love for the American economy to be able to grow in the future somewhere close to the pace it’s grown in the past. Most of the forecasts you see about the growth potential of the economy are at about two percent. This is from the Fed, CBO, maybe even some Wall Street sources. Is the growth potential of the economy at two percent? If it’s going to be more than that, we need stronger productivity growth.

But what is concerning — and what you addressed in your analysis — is the impact of the coronavirus on the productive capacity of the economy. And my concern is that productive capacity will be hurt in such a way that instead of being a two-ish percent economy, maybe we’ll end up being just a one percent economy because of all the damage it’s going to cause. So looking at it now — and it’s still early — let’s tick through a few of the things you’re looking at, or that we should be looking at.

One is labor. You want workers to be productive, they have to be skilled. What will be the impact on the productive capacity of workers, as far as human capital goes?

So I think there are two things we talk about in this study pulling in opposite directions here.

One is that kids are out of school. Now, I know there’s an online replacement for that. My four are all, as we speak, in their “classes.” It’s not clear that’s a perfect substitute. I don’t think it is. And we don’t know how good of a substitute it is. If this goes on for a while, say into the next fall, you could imagine a case where students are effectively missing maybe something like a quarter year, maybe even half a year of schooling because of this.

That might not sound like a lot, but if you go read the literature on educational interventions, if you find something that gives students an extra quarter year of schooling, that’s a huge deal. That is a big, very successful intervention. So something one might worry about going forward is there’s a decline in the amount of schooling that students are getting now as a result of having to be out of the classroom.

Pulling the other way, at the tertiary level, there’s good evidence that schooling is countercyclical. So more people get more training, more college at multiple levels during downturns than booms, because the opportunity cost of their time is falling. And so that’s going to tend to pull the other way, and raise the average human capital level in the economy.

The final, empirical question of which of these two countervailing effects is the biggest will determine whether the crisis has a net positive or negative effect on human capital. But those two things struck us as the most important things to look at coming out of this on the labor side.

It seems to me that there also might be just the learning-by-doing aspect: whether or not we go out and all take classes, maybe the pandemic is forcing everyone who’s working at home to become a little more IT-literate, whether it’s Zoom or other sorts of productivity software. I mean, do you think that’s something significant or potentially significant?

Sure. I think a lot of people have learned how to do their jobs in ways that they didn’t have to worry about before. That is developing human capital. The question going forward is: How applicable is that new knowledge going to be, as we hopefully get back to a pre-pandemic health situation?

So some things will probably persistently change regarding the way we do our jobs, but other things might not. Other things might be stop-gap measures, and they’re not really good substitutes for the way things were done before. In the latter case, people will learn how to do the stop-gap, but it’s not really going to be useful going forward. The former case — where the way things are done in particular kinds of businesses or industries change more persistently and now that training is what people are doing, maybe because they’re forced to — will have more persistent, beneficial, long-run effects.

Well, that’s labor, and then we have capital. And while this pandemic’s terrible, it’s not a war. We’re not seeing the large-scale destruction of physical capital. What is the impact on capital, if anything?

Yeah, so you’re right, we’re not mowing down factories and losing productive capacity that way. This has been mentioned a little bit in the context of, for example, oil extraction — more people are worried about what this is going to mean for that industry. We’re not going to forget where the oil is or how to drill for it and pull it out of the ground. That’s always going to be there.

I think that the effects on capital instead would be that some kinds of capital are probably going to be obsolesced faster than people expected. Let’s say people don’t fly at the same rate they did before the crisis. Well, you’ve got a bunch of planes now that aren’t worth what they were before. Maybe some hotels in far-flung locations, similarly, are not going to yield the economic return one would have expected from them six months ago.

We already were in a low real interest rate environment. We’ve sunk only deeper into lower interest rates. That’s going to make it cheaper to invest in new capital, and you can see where that new capital would come from. The new physical capital is going to be things like health infrastructure and public health robustness, that kind of stuff.

So there’ll be a little bit of reallocation across capital types. It’s not clear though, like you mentioned, that there’s going to be an obvious pull, positive or negative, on the overall level of capital.

Well, I mean, if we’re investing in health infrastructure, does that obviously boost productivity? It very well may if we have another pandemic, but if this is a once-in-a-hundred-years thing, we’re spending a lot of money to deal with something that may not happen again for generations.

Yeah, very good point. Whenever you are spending resources to address these kinds of issues, you might not ever, or not in any reasonable horizon, see any payoff to those investments. It’s been 102 years since the Spanish Flu. If the next coronavirus is 102 years from now, you’re right. Some of those investments we’re making — even though it may be wise along certain dimensions — from a measured productivity point of view, it’s just going to look like a bunch of nonproductive capital.

I was trying to think of another example. I thought of all the spending on technology updates approaching the year 2000. Everyone was worried about the Y2K bug shutting down all the computer systems, and we spent a lot of money to prevent that from happening. But there was also a knock-on effect: We all had an upgraded computer structure. Is that analogy at all relevant for this situation?

Yeah, that’s possible. We talk about that not just in the capital sense, but in the productivity firm level efficiency sense. You could tell a story where a shock like this is such a big thing, that it basically forces a reckoning on the part of companies to think about the way things are done. And that reckoning makes them realize, “Oh, we were doing this thing for this-and-that reason, but that was decided 25 years ago because of something that doesn’t matter anymore. Why are we really doing it this way? Can’t we do it this other way instead?” And you get productivity gains that way.

We know that productivity gains do come from companies’ responses to market conditions. On the demand side, for example, pharmaceutical companies shift the aim of their development efforts based on demographic patterns coming 10, 20 years in the future. On the supply side, we know that companies in high relative wage environments use more capital-intensive technologies, and vice versa for companies in low relative wage environments.

So companies do respond to the market environment around them when figuring out how to produce things. Whether a broad-based, otherwise negative shock is going to induce these kinds of reckoning-driven productivity gains — that’s an open question.

It’s actually been proposed before in the context of environmental regulations. There’s this notion called the Porter hypothesis: Being subject to new environmental regulations could actually make companies more productive for the reasons I was talking about. They’re forced to do things differently, and rather than just changing things on the margin, they reorganize themselves from the ground up in a way that’s hopefully optimal. The evidence is a little bit mixed on that.

But on the other hand, coronavirus is probably a much broader shock to companies than environmental regulations were and are. And so, I think a lot of companies are going to be dealing with: “Why do we do things the way we do? Is there a better way to do it?” You could see productivity gains coming from that process.

One of the things that I’ve been concerned about is the widespread destruction of small- and medium-sized businesses. On one hand, I see it as a bad thing because if we want the recovery to be strong, you want these workers to have some place to go back to. If these businesses disintegrate, you lose not just their connection with employees, but their connection with suppliers and the things that business has learned over the years. You lose all that, and that would be bad for productivity.

But if the result is smaller firms being snapped up by bigger firms, aren’t bigger firms generating a lot more productivity? So again, how do you balance off those two things? The downsides of business destruction versus maybe a little bit more creative destruction?

That’s a great question. The things you mentioned are two different phenomena, but they act in concert in the way you’re describing. So, on the one hand — and we talk about this in the study —there’s a lot of intangible capital that can get destroyed when a company falls apart. And it is more than just the factory, the storefront, whatever it is, the physical capital. You need to go find new workers. Those new workers don’t have relationships with each other or with the company that they did before. You have organizational processes that have been developed, maybe good and bad, but you’re throwing out the good with the bad, and so you lose that intangible capital. And so there is a real cost to the destruction of a company. In an idealized bankruptcy process, you get to keep all that intangible capital as you reorganize, but it’s not clear that it works that easily. So you’ve got that on the one hand.

On the other hand, as you mentioned, it tends to be that larger firms are more efficient. And during normal times, if you see a reallocation towards larger, more efficient firms, well, that raises productivity in the industry. That’s a good thing.

The question is, what’s going to drive the business failures that we’re unfortunately likely to see coming out of this? If it’s highly related to productivity, and the businesses that go out of business are systematically less productive than the ones that stay in business, then the scale is going to tip towards the productivity-enhancing side. If survival is instead determined by other things — like market power, political connections, or who can race to the bank faster for the PPP — that’s not related to productivity. Well, now you’ve just destroyed a bunch of this intangible capital needlessly.

And I think that, empirically, is the key question going forward. What drives business selection and business survival coming out of this? If it’s about productivity, yes, there’ll be some stuff destroyed as businesses go out of business. That’s not great. But on the other hand, you’re reallocating activity towards more efficient producers.

If it’s instead driven by other factors, now it’s a double whammy. You’re destroying all this intangible capital, and you’re not getting any of the productivity gains from reallocation on the other side. And I tell you, that’s going to be one of the first things I look at as soon as I get data on it.

There’s a concern that you have companies, big and small, that were passing the market test as of January, but now they’re in trouble. And well, maybe we should bail those out, because obviously they seemed to be viable businesses three months ago. If not for this, then they’d still be viable businesses.

But then there are companies that were shaky before the crisis, and there seems to be a lot more hesitancy or concern about bailing those companies out. If you bail those companies out, then maybe you really do create the zombie companies that were failing the market test, and are only in existence going forward because of cheap money, whether from the taxpayer or from the Federal Reserve.

I think that’s one of the concerns with a company like Boeing, which seemed to be in a lot of trouble leading into this crisis. And do you bail them out or not — let them go bust? But oh boy, they’re a pretty important American company. There seems to be a number of the concerns around that situation.

I think you’re exactly right. I totally understand that concern. And you framed out the things I was talking about very succinctly. It’s how discerning this selection “sai” is going to be as it cuts through the economy? Is it going to select on things that the market was selecting on before, like efficiency? Or is it going to select on market power, political connections, that kind of thing?

And I don’t want to speak too much towards any specific case. Boeing’s tough because you can imagine there’re a lot of intangibles there. The simple fact is, some companies will go out of business because of this. How discerning is that process going to be?

And I’ll also just mention since we’re on it — if the exit of some of these companies makes room for entry, then you want good entrants. What makes me nervous there is that we’ve been seeing decades-long declines in business formation rates. And maybe that’ll turn around, but certainly the history prior to the coronavirus crisis doesn’t make one encouraged about the notion that you’re going to have a bunch of great new companies starting up on the back end of this thing to replace the ones that go out of business.

Well, indeed. I’m not sure to what degree this has played a role in the lack of business formation, but a concern of mine is risk aversion — we’re going to become the “super cautious society.” Consumers aren’t going to want to spend, businesses aren’t going to want to invest, government will suddenly become worried about deficits again.

And maybe, most importantly, people will be more reluctant to give it a go and start a business. Whether it’s a smaller, more mom-and pop-business, or something that actually could turn to a really high impact company down the road. Should I be concerned about risk aversion becoming a bigger factor in the American economy?

That is something I’ve been thinking more about lately. Other people have raised this over the past several years. This is kind of related to Marc Andreessen’s recent essay, “It’s Time to Build.” Why don’t we build things the way we used to? And this also overlaps with a bit of the context about the productivity slowdowns of recent and maybe even prior decades.

The evidence on this is still not well-developed, but I think it is something to be worried about and to keep an eye on, because just all sorts of measures of dynamism — for lack of a better word — have been falling. I mentioned business formation, but also just job turnovers and the geographic mobility of American workers has been falling for decades. Maybe that’s in part tied to this risk aversion thing. And if this just makes it worse and exacerbates it, yeah, that’s not going to be a good thing. So, it is something I would worry about and want to keep an eye on.

My guest today has been Chad Syverson. Chad, thanks for coming on the podcast.

My pleasure, thanks! It was fun.